The Protocol
Five principles first gathered at REMIX Sydney. They are a starting point, not a standard. Each is paired with a behaviour you can test and a condition for stopping. Participant responses are synthesised by the author; they are never automatically published here.
- 1
Pause and consult the mesh
Testable behaviourBefore relying on an AI answer that affects others, check it with at least one person, place or practice that holds relevant knowledge.
Pause conditionPause when the answer arrives faster than you can name who else should be asked.
- 2
Protect embodied learning
Testable behaviourKeep tasks that build skill in the hands and body — drawing, writing by hand, fieldwork — free of AI in learning settings.
Pause conditionPause when the tool begins to replace the practice it was meant to support.
- 3
Keep authorship visible
Testable behaviourLabel AI-assisted passages, images or decisions where they are shared, and credit the people and communities whose knowledge shaped them.
Pause conditionPause when you can no longer say which parts are yours.
- 4
Use AI at an earned scale
Testable behaviourMatch the size of the model and the energy it uses to the actual need; prefer smaller, local or no tools where they suffice.
Pause conditionPause when the computation outweighs what is gained.
- 5
Close and look up
Testable behaviourEnd sessions deliberately and return attention to the room, the people and the place you are in.
Pause conditionClose the tab when the conversation with the machine is crowding out the one beside you.
Every revision is recorded here with its source.
- v0.12026 · REMIX Sydney
Five principles drafted from the Collective Care Protocol Lab worksheets.
- v0.2Planned · after REMIX Perth
Revisions to be synthesised by hand from Perth contributions. Not yet published.